“Innovation has nothing to do with how many R&D dollars you have. When Apple came up with the Mac, IBM was spending at least 100 times more on R&D. It's not about money. It's about the people you have, how you're led, and how much you get it.” - Steve Jobs

Ontology isn’t overrated and Tagging isn’t really anything new.

Ontology is Overrated: Categories, Links, and Tags by Clay Shirky is a great article that describes the power of tagging, but it is not critical enough, nor does it separate blatant individualistic tagging from a more controlled universe of folksonomy. Thomas Vander Wal was a guest contributore on Ok-Cancel with an article of his own, Tagging for Fun and Finding.

Is all this lauding of tagging going a bit too far? It all feels short-sighted to me. Fun! but short-sighted.

I would be interested in what others think about all this lauding of Tags, or unstructured metadata? Now, I like the whole idea that I can “tag” things (BTW, the OK-Cancel comic is brilliant) so that *I* can refind things later. But what gets me is the lack of frame of reference, or stable POV.

More than any other group we, user experience professionals, should understand where tagging fails to increase findability:
User Experience
Information Architecture
Interaction Design
Usability (w/ or w/o) Engineering
Experience Design

I may use one or none of the above to label my blog articles, or come up w/ something completely new. There is no shared cultural value for querying the system.

I know … The system will fall into place through organic use, right? Well it hasn’t yet, as I know so far that tagging doesn’t really speak well to the masses, outside of a core select few. Delicious is interesting as is flickr, but when I try to use it, I feel totally overwhelmed with way more than I wanted, and not enough of what I did want.

Now, that being said, like Clay I do agree that the current systems of hierarchy and file systems don’t make sense, but why does having a controlled vocabulary have to mean that it is a hierarchical taxonomy like the Yahoo example he gives. Unlike library catalog systems which tend to want to put an object in one and only one space, the web DOES afford us a non-linear and multiplicity of locations and that’s good, but doesn’t equal the need for uncontrolled vocabularies.

If we see any trend is that we need to have controlled vocabularies that can learn and adjust over time. But unintelligent tagging systems, IMHO, do not increase findability of things that I haven’t found already, just re-findability of those items that I’ve already found.

What is much better is intelligences that create relevancy based on context of use, referencing, and linking. This can be informed by individual tagging, but can’t replace it, nor can indivdiual tagging succeed long term on its own. Otherwise we will end up just where we are today. A big web w/ no real frame of reference.

Be Sociable, Share!
  • Thomas Vander Wal

    While general tagging does not have a point of reference a folksonomy does in the individual person, their tags and the objects they are tagging. In some long discussions with the founder of CiteULike (an academic folksonomy tool similar to del.icio.us) he has been running some insane research on his site and del.icio.us and finding the folksonomies resulting in using these three points of reference provide results he had not been able to get even close to with his taxonomy that he had spent a lot of of time building (he is a chemical biologist by trianing and still working toward his PhD.). He built the tool for himself and has been utterly suprised by the research he has done and the greatly improved results he gets with the tag based algorithems rather than the taxonomy.

    There is also a lot of academic research coming out in the next six months or so showing much improved narrowing of results using folksonomies than for traditional taxonomies/ontologies. This is similar to Jon Udell’s queries he has been running on del.icio.us using people who have tagged objects he tagged with the same tags he did so to find new information. The results he is getting are quite strong, much better than Google or Technorati can do at this point.

    This is rather stunning considering del.icio.us has orders of magnitude fewer entries than Google, or even Technorati. I know Google and Yahoo are investing heavily in folksonomies and social tagging. Google is trying to improve upon their vast improvement over other search engines in 1998 by adding the consumer of informaiton’s annotations in their algorithems (what is in the anchor tag). I know the research that these two companies have been doing has been providing greatly improved search capabilities. This goes counter to the quote “use Google to find what you know you want to find and use del.icio.us for what you did not know you wanted to find”, but it is working for these groups that have figured out how to use the points of reference in folksonomies to theirm benefit.

  • Dave

    Hi Thomas,
    Great post … thanx! I don’t think we are that far off. By using a human reference point, you are in essence creating the controlled vocabulary I was speaking of. That person’s vacabulary is what you will be working with. But for this to work, it means someone needs to know that that person exists AND have a relationship or at least knowledge point with him/her in the first place.

    I think that Tagging has value for the individual, in terms of re-findability, but I cannot see how tagging by itself remedies much of the other findability issues we are facing today with so much information.

    Now, we could hypothesize that tagging can lead to new vocabularly sets over “generations” of tagging. Meaning, a lot of trial and error is going to have to take place before there is any level of consistency across a mass of individuals. I’ll be interested to see how all this changes over time. I do think again that a living intelligent controlled vocabularly is the key to this issue. Tagging will inform it, but it won’t be it. Until the search engine knows that UX = user experience no matter who tags it, I think we will have a while to go before this all pays off.


The archives run deep. Feel free to search older content using topic keywords.